net.tomp2p.rpc

Class SimpleBloomFilter<E>

java.lang.Object

net.tomp2p.rpc.SimpleBloomFilter<E>

Type Parameters:

E - The type of object the BloomFilter should contain

All Implemented Interfaces:

Serializable, Iterable<E>, Collection<E>, Set<E>

public class SimpleBloomFilter<E>
extends Object
implements Set<E>, Serializable

A simple Bloom Filter (see http://en.wikipedia.org/wiki/Bloom_filter) that uses java.util.Random as a primitive hash function, and which implements Java's Set interface for convenience. Only the add(), addAll(), contains(), and containsAll() methods are implemented. Calling any other method will yield an UnsupportedOperationException. This code may be used, modified, and redistributed provided that the author tag below remains intact.

Author:

Ian Clarke , Thomas Bocek Added methods to get and create a SimpleBloomFilter from existing data. The data can be either a BitSet or a bye[].

See Also:

Serialized Form

Field Summary

Fields

Modifier and Type	Field and Description
static int	SIZE_HEADER

Constructor Summary

Constructors

Constructor and Description
SimpleBloomFilter(io.netty.buffer.ByteBuf channelBuffer) Constructs a SimpleBloomFilter out of existing data.
SimpleBloomFilter(double falsePositiveProbability, int expectedElements)
SimpleBloomFilter(int byteArraySize, int expectedElements) Construct an empty SimpleBloomFilter.
SimpleBloomFilter(int byteArraySize, int expectedElements, BitSet bitSet) Constructs a SimpleBloomFilter out of existing data.

Method Summary

Methods

Modifier and Type	Method and Description
boolean	add(E o)
boolean	addAll(Collection<? extends E> c)
void	clear() Clear the Bloom Filter.
boolean	contains(Object o)
boolean	containsAll(Collection<?> c)
boolean	equals(Object obj)
int	expectedElements() Returns the expected elements that was provided by the user.
double	expectedFalsePositiveProbability() Calculates the approximate probability of the contains() method returning true for an object that had not previously been inserted into the bloom filter.
BitSet	getBitSet() Returns the bitset that backs the bloom filter.
int	hashCode()
boolean	isEmpty() Not implemented.
Iterator<E>	iterator() Not implemented.
SimpleBloomFilter<E>	merge(SimpleBloomFilter<E> toMerge)
SimpleBloomFilter<Number160>	not() Invert the bloom filter
boolean	remove(Object o) Not implemented.
boolean	removeAll(Collection<?> c) Not implemented.
boolean	retainAll(Collection<?> c) Not implemented.
int	size() Not implemented.
Object[]	toArray() Not implemented.
<T> T[]	toArray(T[] a) Not implemented.
void	toByteBuf(io.netty.buffer.ByteBuf buf) Converts data to a byte buffer.
String	toString()

Methods inherited from class java.lang.Object

clone, finalize, getClass, notify, notifyAll, wait, wait, wait

Field Detail

SIZE_HEADER

public static final int SIZE_HEADER

See Also:

Constant Field Values

Constructor Detail

SimpleBloomFilter

public SimpleBloomFilter(int byteArraySize,
                 int expectedElements)

Construct an empty SimpleBloomFilter. You must specify the number of bits in the Bloom Filter, and also you should specify the number of items you expect to add. The latter is used to choose some optimal internal values to minimize the false-positive rate (which can be estimated with expectedFalsePositiveRate()).

Parameters:

byteArraySize - The number of bits in multiple of 8 in the bit array (often called 'm' in the context of bloom filters).

expectedElements - The typical number of items you expect to be added to the SimpleBloomFilter (often called 'n').

SimpleBloomFilter

public SimpleBloomFilter(double falsePositiveProbability,
                 int expectedElements)

SimpleBloomFilter

public SimpleBloomFilter(io.netty.buffer.ByteBuf channelBuffer)

Constructs a SimpleBloomFilter out of existing data.

Parameters:

channelBuffer - The byte buffer with the data

SimpleBloomFilter

public SimpleBloomFilter(int byteArraySize,
                 int expectedElements,
                 BitSet bitSet)

Constructs a SimpleBloomFilter out of existing data. You must specify the number of bits in the Bloom Filter, and also you should specify the number of items you expect to add. The latter is used to choose some optimal internal values to minimize the false-positive rate (which can be estimated with expectedFalsePositiveRate()).

Parameters:

byteArraySize - The number of bits in multiple of 8 in the bit array (often called 'm' in the context of bloom filters).

expectedElements - he typical number of items you expect to be added to the SimpleBloomFilter (often called 'n').

bitSet - The data that will be used in the backing BitSet

Method Detail

expectedFalsePositiveProbability

public double expectedFalsePositiveProbability()

Calculates the approximate probability of the contains() method returning true for an object that had not previously been inserted into the bloom filter. This is known as the "false positive probability".

Returns:

The estimated false positive rate

expectedElements

public int expectedElements()

Returns the expected elements that was provided by the user.

Returns:

The expected elements that was provided by the user

add

public boolean add(E o)

Specified by:

add in interface Collection<E>

Specified by:

add in interface Set<E>

Parameters:

o - Add element

Returns:

This method will always return false

See Also:

Set.add(java.lang.Object)

addAll

public boolean addAll(Collection<? extends E> c)

Specified by:

addAll in interface Collection<E>

Specified by:

addAll in interface Set<E>

Parameters:

c - The elements to add

Returns:

This method will always return false

clear

public void clear()

Clear the Bloom Filter.

Specified by:

clear in interface Collection<E>

Specified by:

clear in interface Set<E>

contains

public boolean contains(Object o)

Specified by:

contains in interface Collection<E>

Specified by:

contains in interface Set<E>

Parameters:

o - The object to compare

Returns:

False indicates that o was definitely not added to this Bloom Filter, true indicates that it probably was. The probability can be estimated using the expectedFalsePositiveProbability() method.

containsAll

public boolean containsAll(Collection<?> c)

Specified by:

containsAll in interface Collection<E>

Specified by:

containsAll in interface Set<E>

Parameters:

c - The collection to check

Returns:

true if all elements of the collection are in this bloom filter

isEmpty

public boolean isEmpty()

Not implemented.

Specified by:

isEmpty in interface Collection<E>

Specified by:

isEmpty in interface Set<E>

Returns:

nothing

iterator

public Iterator<E> iterator()

Not implemented.

Specified by:

iterator in interface Iterable<E>

Specified by:

iterator in interface Collection<E>

Specified by:

iterator in interface Set<E>

Returns:

nothing

remove

public boolean remove(Object o)

Not implemented.

Specified by:

remove in interface Collection<E>

Specified by:

remove in interface Set<E>

Parameters:

o - nothing

Returns:

nothing

removeAll

public boolean removeAll(Collection<?> c)

Not implemented.

Specified by:

removeAll in interface Collection<E>

Specified by:

removeAll in interface Set<E>

Parameters:

c - nothing

Returns:

nothing

retainAll

public boolean retainAll(Collection<?> c)

Not implemented.

Specified by:

retainAll in interface Collection<E>

Specified by:

retainAll in interface Set<E>

Parameters:

c - nothing

Returns:

nothing

size

public int size()

Not implemented.

Specified by:

size in interface Collection<E>

Specified by:

size in interface Set<E>

Returns:

nothing

toArray

public Object[] toArray()

Not implemented.

Specified by:

toArray in interface Collection<E>

Specified by:

toArray in interface Set<E>

Returns:

nothing

toArray

public <T> T[] toArray(T[] a)

Not implemented.

Specified by:

toArray in interface Collection<E>

Specified by:

toArray in interface Set<E>

Type Parameters:

T - nothing

Parameters:

a - nothing

Returns:

nothing

getBitSet

public BitSet getBitSet()

Returns the bitset that backs the bloom filter.

Returns:

bloom filter as a bitset

toByteBuf

public void toByteBuf(io.netty.buffer.ByteBuf buf)

Converts data to a byte buffer. The first two bytes contain the size of this simple bloom filter. Thus, the bloom filter can only be of length 65536.

Parameters:

buf - The byte buffer where the bloom filter will be written.

merge

public SimpleBloomFilter<E> merge(SimpleBloomFilter<E> toMerge)

Parameters:

toMerge - Merge to bloom filters using OR

Returns:

A new bloom filter that contains both sets.

equals

public boolean equals(Object obj)

Specified by:

equals in interface Collection<E>

Specified by:

equals in interface Set<E>

Overrides:

equals in class Object

hashCode

public int hashCode()

Specified by:

hashCode in interface Collection<E>

Specified by:

hashCode in interface Set<E>

Overrides:

hashCode in class Object

toString

public String toString()

Overrides:

toString in class Object

not

public SimpleBloomFilter<Number160> not()

Invert the bloom filter